The Role of the Prefrontal Cortex in the Expression of Impulsive- and

Premeditated-Aggression

by

Sarah Haberle BA (Hons)

Submitted in fulfilment of the requirements for the Degree of Doctor of Philosophy

(Clinical Psychology)

School of Psychology

University of Tasmania

October 2011

i

This thesis contains no material, which has been accepted for a degree or

diploma by the University or any other institution, except by way of background

information and duly acknowledged in the thesis. To the best of my knowledge and

belief, this thesis contains no material previously published or written by another

person except where due acknowledgement is made in the text of the thesis, nor does

the thesis contain any material that infringes copyright.

__________________________________ Date: _______________

Sarah Haberle

The research associated with this thesis abides by the international and

Australian codes on human and animal experimentation, the guidelines by the

Australian Government’s office of the Gene Technology Regulator and the rulings of

the Safety, Ethics and Institutional Biosafety Committees of the University.

__________________________________ Date: _______________

Sarah Haberle

This thesis may be made available for loan and limited copying in accordance

with the Copyright Act 1968.

__________________________________ Date: _______________

Sarah Haberle

ii

Abstract

The notion that there is a relationship between frontal lobe damage and

aggressive behaviour has been recognised in the clinical literature for over 50 years.

However, although there is evidence for an association between general brain

dysfunction and aggression, there is little evidence pertaining to subclinical

impairment and the propensity for aggressive behaviour. Further to this, given the

functionally heterogeneity of the prefrontal cortex, it is vital to delineate the specific

roles of the dorsolateral, orbitofrontal and medial aspects of the prefrontal cortex in

the expression of aggression.

Two forms of aggression are distinguished: reactive, impulsive-aggression

and goal-directed premeditated aggression. While impulsive-aggression is typically

described as an emotionally-charged aggressive response characterised by a lack of

control, premeditated aggression is considered to be a planned and controlled

aggressive display that is instrumental in nature. The qualitative differences between

these subtypes of aggression suggest distinct neuropsychological differences

mediating the likelihood of their display.

The aim of this thesis was to clarify the role of the prefrontal cortex in

subclinical impulsive-aggression and premeditated aggression. More specifically,

possible executive functioning deficits mediated by the dorsolateral prefrontal cortex,

and emotion recognition, impulsivity, and response reversal capabilities mediated by

the orbitofrontal cortex were explored. Participants included university

undergraduate students identified as having high levels of trait aggression, classified

as either predominantly impulsive, or predominantly premeditated in nature.

Experiment 1 (n=85) explored possible executive deficits using a battery of

iii

neuropsychological measures pertaining to dorsolateral functioning. It was found that

impulsive-aggressive individuals performed significantly poorer on measures of

cognitive flexibility, planning, problem-solving, and flexibility of verbal thought

processes.

Experiment 2 (n=87) sought to identify possible deficits in interpretations of

facial expressions of emotion and hostile attribution biases. Contrary to expectations,

the results indicated that while impulsive- and premeditated-aggressive individuals

do not incorrectly interpret emotional expressions, premeditated-aggressive

individuals attributed greater levels of aggression to neutral faces.

Experiment 3 (n=87) investigated functions of the orbitofrontal cortex,

namely impulsivity, response reversal, and decision-making capabilities. No

differences between impulsive-aggressive and premeditated-aggressive individuals

were found on any of these measures suggesting negligible involvement of the

orbitofrontal cortex in subclinical aggression.

Overall, the results from this thesis suggest distinct neuropsychological

processes in individuals who display predominantly impulsive-aggressive behaviour

compared to those who display predominantly premeditated-aggression. While

impulsive-aggression may result from executive dysfunction pertaining to the

dorsolateral region of the prefrontal cortex, the display of premeditated aggression is

related to functioning of the orbitofrontal cortex mediating the interpretation of

aggression in others. Such findings have important implications not only in the

understanding of the causal features of such behaviour, but also in the development

and implementation of successful treatment strategies.

iv

Acknowledgements

I would like to express my thanks to friends and colleagues whose support and

encouragement have made this thesis possible. Most importantly, I would like to

thank my supervisor Dr Frances Martin for her continued help and support throughout

my candidature. I greatly appreciate your patience and flexibility, your honesty, and

your continued encouragement. I would also like to thank Dr Raimondo Bruno, Dr

Clive Skilbeck, and Dr Tess Crawley for their valued feedback on my research and

other academic and support staff within the School of Psychology at the University of

Tasmania, particularly Vlasti Broucek and David Chadderton for IT and technical

support, and Sue Ross and Susan Jopling for answering all my questions and helping

out in every way they possibly could. Thank you to all members of the annexe,

especially my roomies Megan Laugher and Matt Treeby for the many laughs, songs,

murals, mega-cards, and dances over the years. Lastly, I would like thank all of my

friends and family for their patience, advice, and understanding, and being so

supportive and encouraging especially when I needed it most.

v

Table of Contents

Abstract Acknowledgements List of Tables List of Figures Chapter 1. Overview of the Thesis p. 1 Chapter 2. Impulsive- and Premeditated-Aggression: A Review of

the Literature p. 6

2.1 Aggression p. 6 2.2 Distinguishing between impulsive- and premeditated-

aggression p. 7 2.2.1 Psychobiological evidence p. 11 2.2.2 Psychophysiological evidence p. 12 2.2.3 Neuropsychological evidence p. 14 2.2.4 Psychopathy p. 15

2.3 Importance of distinguishing between impulsive- and premeditated-aggression p. 17

Chapter 3. Prefrontal Cortex & Aggression p. 20

3.1 The prefrontal cortex p. 22 3.2 Prefrontal divisions p. 24 3.3 Prefrontal dysfunction and aggression p. 28

3.3.1 Lesion studies p. 28 3.3.2 Neuroimaging and clinical neurological findings p. 32

3.4 The specific roles of the orbitofrontal and dorsolateral prefrontal cortex in aggression p. 35

3.5 The role of the prefrontal cortex in impulsive- and premeditated-aggression p. 37

3.6 Subcortical structures p. 40 3.7 Conclusion p. 41

Chapter 4. Rationale p. 42 Chapter 5. Study 1: Executive Functioning p. 46

5.1 Neuroanatomy of executive functions p. 47 5.2 Executive functioning measures p. 50

5.2.1 Verbal Fluency Test p. 50 5.2.2 Trail Making Test p. 51 5.2.3 Tower of Hanoi p. 52 5.2.4 Stroop Colour-Word Interference Task p. 53

vi

5.2.5 The Brixton Test p. 54 5.3 Executive functioning and aggression – a review of previous

research p. 55 5.4 Executive functioning and impulsive- and premeditated-

Aggression p. 57 5.5 The relationship between executive functioning deficits and

aggression p. 60 5.6 Limitations of other studies p. 62 5.7 Aim and hypotheses p. 64 5.8 Method p. 65

5.8.1 Participants p. 65 5.8.2 Materials p. 70

5.8.2.1 Questionnaires p. 70 5.8.2.2 Executive function measures p. 73 5.8.2.3 Wechsler Adult Intelligence Scale –

Third Edition p. 77 5.8.3 Procedure p. 78

5.9 Results p. 80 5.9.1 Participants p. 80 5.9.2 Executive function measures p. 82 5.9.3 Wechsler Adult Intelligence Scale – Third Edition p. 83

5.10 Discussion p. 85 5.10.1 Verbal Fluency Test p. 87 5.10.2 Trail Making Test p. 88 5.10.3 Tower of Hanoi p. 90 5.10.4 Stroop Colour-Word Interference Task p. 92 5.10.5 The Brixton Test p. 95 5.10.6 Personality measures p. 96 5.10.7 Psychopathy and Antisocial Personality Disorder p. 98 5.10.8 The link between executive functioning deficits and

impulsive-aggression p. 100 5.10.9 Conclusion p. 103

Chapter 6. Study 2: Emotion Recognition and Aggression Attribution p. 107

6.1 Neural systems involved in emotion recognition p. 109 6.2 Separable neural systems for different emotional expressions p. 113 6.3 The relationship between emotion recognition and aggression p. 115 6.4 Aim and hypotheses p. 124 6.5 Method p. 125

6.5.1 Participants p. 125 6.5.2 Materials p. 127

6.5.2.1 Questionnaires p. 127 6.5.2.2 Facial recognition tasks p. 128 6.5.2.3 Wechsler Adult Intelligence Scale –

Third Edition p. 129 6.5.3 Procedure p. 129

6.6 Results p. 131 6.6.1 Participants p. 131 6.6.2 Emotion recognition task p. 132

vii

6.6.3 Aggression rating task p. 142 6.6.4 Wechsler Adult Intelligence Scale – Third Edition p. 150

6.7 Discussion p. 150 6.7.1 Emotion recognition task p. 150 6.7.2 Aggression rating task p. 155 6.7.3 Personality measures p. 157 6.7.4 The link between emotion recognition and aggression p. 157 6.7.5 Conclusion p. 159

Chapter 7. Study 3: Inhibition, Response Reversal, and

Decision-Making p. 161 7.1 Inhibition p. 161

7.1.1 Inhibition and the frontal lobes p. 164 7.1.2 Inhibition measures p. 166 7.1.3 Inhibition and aggression p. 169

7.2 Response reversal p. 171 7.2.1 Response reversal and the frontal lobes p. 172 7.2.2 Response reversal measures p. 174 7.2.3 Response reversal and aggression p. 177

7.3 Decision-making p. 180 7.3.1 Decision-making and the frontal lobes p. 181 7.3.2 Decision-making measures p. 181 7.3.3 Decision-making and aggression p. 184 7.3.4 Somatic marker hypothesis p. 186

7.4 Impulsivity, response reversal, decision-making, and aggression p. 187 7.5 Aim and hypotheses p. 188 7.6 Method p. 189

7.6.1 Participants p. 189 7.6.2 Materials p. 189

7.6.2.1 Questionnaires p. 189 7.6.2.2 Inhibition, response reversal, and

decision-making tasks p. 189 7.6.2.3 Wechsler Adult Intelligence Scale –

Third Edition p. 191 7.6.3 Procedure p. 191

7.7 Results p. 195 7.7.1 Stop Signal Task p. 195 7.7.2 Intra/Extra Dimensional Set Shift task p. 196 7.7.3 Cambridge Gambling Task p. 198

7.8 Discussion p. 202 7.8.1 Stop Signal Task p. 203 7.8.2 Intra/Extra Dimensional Set Shift task p. 207 7.8.3 Cambridge Gambling Task p. 210 7.8.4 Limitations and directions for future research p. 212 7.8.5 Conclusion p. 213

viii

Chapter 8. General Discussion p. 215 8.1 Overview of findings p. 215

8.1.1 Executive functioning p. 215 8.1.2 Emotion recognition and aggression attribution p. 218 8.1.3 Inhibition, response reversal, and decision-making p. 219

8.2 Theoretical implications p. 220 8.2.1 Inhibition p. 222 8.2.2 Premeditated-aggression p. 224

8.3 Clinical implications p. 226 8.4 Limitations p. 228 8.5 Conclusion p. 229

References p. 231 Appendices p. 307

Appendix A: HREC approval letter p. 308 Appendix B: Information sheet for Study 1 p. 310 Appendix C: Consent form for Study 1 p. 312 Appendix D: Facial stimuli from Ekman and Friesen’s (1976)

collection used for the emotion recognition task p. 313 Appendix E: Facial stimuli from Ekman and Friesen’s (1976)

collection used for the aggression rating task p. 314 Appendix F: Information sheet for Study 2 and Study 3 p. 315 Appendix G: Consent form for Study 2 and Study 3 p. 317 Appendix H: Data analyses CD

ix

List of Tables

Chapter 5

Study 1: Executive Functioning

Table 5.1 The Impulsive-Premeditated Aggression Scale p. 67

Table 5.2 Number of males and females in the three participant

groups and total sample p. 68

Table 5.3 Mean (and standard deviation) scores on the Aggression

Questionnaire – Short Form and ages for the three

participant groups and total sample p. 69

Table 5.4 Means (and standard deviations) for the Aggression

Questionnaire – Full Scale and I7 Impulsivity

Questionnaire for the three participant groups p. 81

Table 5.5 Means (and standard deviations) for the three participant

groups on the executive function and WAIS-III measures p. 84

Table 5.6 Results of ANOVAs for the executive function and

WAIS-III measures p. 85

Chapter 6

Study 2: Emotion Recognition and Aggression Attribution

Table 6.1 Number of males and females in the three participant

groups and total sample p. 126

Table 6.2 Mean (and standard deviations) scores on the Aggression

Questionnaire – Short Form and ages for the three

participant groups and total sample p. 127

Table 6.3 Means (and standard deviations) for the subscales of the

Aggression Questionnaire – Full Scale and I7 Impulsivity

Questionnaire for the three participant groups p. 132

Table 6.4 Mean (and standard deviations) number of correct

responses (maximum = 4) on the emotion recognition task

for the three participant groups p. 133

x

Table 6.5 Paired samples t-test results for response for each face

type p. 135

Table 6.6 Mean (and standard deviations) frequency of responses to

the neutral face on the emotion recognition task for the

three participant groups p. 137

Table 6.7 ANOVA results for responses to the neutral face for the

main effect of participant group at the 1000ms and 2000ms

stimulus durations p. 138

Table 6.8 Paired samples t-test results for frequency of response for

each face type when interpreting the neutral face p. 139

Table 6.9 Mean (and standard deviations) reaction times on the

emotion recognition task for the three participant groups p. 140-141

Table 6.10 Paired samples t-test results for reaction time for each

face type p. 142

Table 6.11 Mean (and standard deviations) responses for the three

participant groups on the aggressive rating task p. 144

Table 6.12 Mean (and standard deviations) reaction times for the

three participant groups on the aggressive rating task p. 145

Table 6.13 Paired samples t-test results for face type at the 1000ms

stimulus duration p. 148

Table 6.14 Paired samples t-test results for face type at the 2000ms

stimulus duration p. 149

Table 6.15 Means (and standard deviations) for Vocabulary and

Digit Span for the three participant groups p. 150

Chapter 7

Study 3: Inhibition, Response Reversal, and Decision-Making

Table 7.1 Means (and standard deviations) for the Stop Signal Task

for the three participant groups p. 196

Table 7.2 Means (and standard deviations) on the Intra-Extra

Dimensional Set Shift task for the three participant groups p. 198

Table 7.3 Results of univariate ANOVAs for the Intra-Extra

Dimensional Set Shift task p. 199

xi

Table 7.4 Means (and standard deviations) on the Cambridge

Gambling Task for the three participant groups p. 201

Table 7.5 Results of univariate ANOVAs for the Cambridge

Gambling Task p. 202

xii

List of Figures

Chapter 6

Study 2: Emotion Recognition and Aggression Attribution

Figure 6.1 Mean number of correct responses (maximum = 4) for each

face type at the 1000ms and 2000ms stimulus durations p. 136

Figure 6.2 Reaction times for each face type by each group at the

1000ms stimulus duration p. 147

Figure 6.3 Reaction times for each face type by each group at the

2000ms stimulus duration p. 149

1

Chapter 1

Overview of the Thesis

While the underpinnings of human aggression are clearly multifactorial, including

political, socioeconomic, cultural, and psychological factors, it is also clear that some

forms of aggression, either impulsive or premeditated in nature, have an underlying

neurobiology that is only just beginning to be understood. In this research the

neurobiology of aggression is addressed, specifically the role of the prefrontal cortex

in the expression of both impulsive- and premeditated-aggression.

A significant body of evidence indicates that the likelihood of acting

aggressively is related to the functional capacity of the frontal lobe. Using

neuroimaging techniques, studies of violent offenders have consistently shown

abnormalities in frontal lobe structures in individuals who have histories of violence

(e.g., Raine, Lencz, Bihrle, LaCasse & Colletti, 2000; Raine et al., 1998).

Additionally, lesion studies (e.g., Damasio, Grabowski, Frank, Galaburda & Damasio,

1994) and neuropsychological studies (e.g., Stanford, Greve & Gerstle, 1997) have

provided evidence of the relationship between prefrontal impairment and the

propensity for aggressive behaviour. Unfortunately, however, the above studies have

placed little emphasis on the separable regions of the prefrontal cortex. This is despite

the fact that neuropsychological data strongly suggesting that only medial and

orbitofrontal regions of the prefrontal cortex are involved in mediating aggression,

while dorsolateral prefrontal cortex has little role (Grafman et al., 1996). The

prefrontal cortex is functionally and anatomically heterogeneous and thus the

separable regions may be differentially involved in the expression of aggression. The

present thesis thus attempts to delineate the specific roles of these subregions in

impulsive- and premeditated-aggression.

2

Many researchers have suggested that the relationship between prefrontal

abnormalities and likelihood of aggression is mediated by the failure to adaptively use

executive functions (Giancola, 2000). As outlined in Chapter 5, executive functions is

a broad term used to describe those abilities which allow an individual to respond to

situations in a flexible manner, creating and adapting plans, and not being governed

exclusively by external stimuli (Hoaken, Allaby, & Earle, 2007). Such abilities are

presumed to be mediated predominantly by the dorsolateral region of the prefrontal

cortex.

A further ability linked to the prefrontal cortex is the ability to correctly

interpret emotional facial expressions. More specifically, patients with orbitofrontal

cortex lesions are impaired in their ability to recognise facial expressions, particularly

anger (Blair & Cipolotti, 2000; Hornak, Rolls & Wade, 1996). Neuroimaging studies

support these findings, demonstrating activation in the orbitofrontal cortex by

negative emotional expressions; in particular, anger, but also fear and disgust (Blair,

Morris, Frith, Perrett & Dolan, 1999; Kesler-West et al., 2001). As described in

Chapter 6, aberrations in the ability to identify facial expressions may result in the

generation of inappropriate social responses, such as reacting aggressively to

ambiguous social situations (Dodge, Laird, Lochman & Zelli, 2002). This hypothesis

is based on Dodge (1986) who proposed that the accurate interpretation of social

stimuli must be completed for prosocial behaviour to be manifested.

Individuals with lesions to the prefrontal cortex have also been shown to have

deficits in inhibition, decision-making, and response reversal (Bechara, Damasio,

Damasio & Anderson, 1994; Rolls, Hornak, Wade & McGrath, 1994). Such patients

also display impairment in social functioning. The suggestion then follows that

reappraisal will play an important role in social contexts in which one is required to

3

adapt to rapidly changing contexts (Happaney, Zelazo & Stuss, 2004). Furthermore,

an inability to suppress previously rewarded responses due to inhibitory deficits, will

in turn lead to inappropriate social responses.

In research on aggression, it is vital to distinguish between impulsive- and

premeditated-aggression. Impulsive-aggression is more reactive in nature and

displayed without a self-generated goal. Premeditated-aggression, in contrast, appears

to occur without provocation, is proactive, and is seen as a means to gain a valued

outcome. This heterogeneity between impulsive- and premeditated-aggression

suggests distinct cognitive mechanisms responsible for their display.

Both the animal and human neuropsychological literature suggests that the

prefrontal cortex is involved in the modulation of impulsive-aggression (Anderson,

Bechara, Damasio, Tranel & Damasio, 1999; Gregg & Siegel, 2001). Certainly,

damage to the medial frontal and orbitofrontal cortex is associated with increased risk

for the display of impulsive-aggression in humans whether the lesion occurs in

childhood (Anderson et al., 1999) or adulthood (Grafman et al., 1996). More

specifically, individuals with orbitofrontal cortex lesions are typically described as

disinhibited, socially inappropriate, impulsive, irresponsible, and as often

misinterpreting others‟ moods (Rolls et al., 1994). In addition, there are considerable

neuroimaging data assessing the neural functioning of patients with impulsive-

aggression. These data have revealed reduced prefrontal functioning in patients

presenting with impulsive-aggression (Søderstrom, Tullberg, Wikkelso, Ekholm &

Forsman, 2000). Interestingly, this reduced prefrontal functioning is not observed in

those with predominantly premeditated-aggression (Raine et al., 1998).

Accounts of premeditated-aggression often lie in the related construct of

psychopathy. Psychopathic individuals, individuals who present with marked

4

premeditated-aggression, do not present with deficits on neuropsychological measures

which pertain predominantly to the dorsolateral prefrontal cortex (Mitchell, Colledge,

Leonard & Blair, 2002). Psychopathic individuals‟ high level of premeditated-

aggression is thus completely unlike that of patients with orbitofrontal cortex lesions

(Cornell et al., 1996). It is likely then that such individuals show elevated levels of

premeditated-aggression because they have been reinforced, and not punished, for

committing such behaviour in the past (Blair, 2004). Such aversive conditioning has

been shown to be mediated by the amygdala. In support of this, an MRI study by

Tiihonen et al. (2000) found a strong negative correlation between level of

psychopathy and amygdala volume.

However, while there is clear evidence of amygdala involvement in

premeditated-aggression through its role in aversive conditioning and instrumental

learning, the orbitofrontal cortex may also be involved through its role in response

reversal and extinction. That is, changing a response to a stimulus when the

reinforcement contingencies change (Dias, Robbins & Roberts, 1996a). Moreover, the

orbitofrontal cortex has been linked to decision-making when knowledge about

potential positive and negative results is necessary to guide behavioural responding

(Rogers et al., 1999b). On tasks such as the Intradimensional/Extradimensional

(ID/ED) Set Shift Task which involves response reversal, adult psychopathic

individuals show impairment (Mitchell et al., 2002). This suggests possible

orbitofrontal dysfunction in individuals presenting with marked premeditated

aggression.

Investigating prefrontal impairment in antisocial behaviour more broadly is

problematic due to its heterogeneity and comorbidity with several disorders, including

drug and alcohol abuse, Antisocial Personality Disorder (APD), pathological

5

gambling, schizophrenia, and bipolar disorders, all of which may or may not involve

an aggressive component. Frontal lobe functions have been implicated in all of these

comorbid conditions, however frontal lobe deficits have also been found in disorders

not readily associated with antisocial behaviour, such as obsessive-compulsive

disorder. The question thus lies in whether there is a single common component of

these disorders given that studies on such populations have been inconsistent with

regard to demonstrated neuropsychological deficits.

Through an investigation of prefrontal functioning in impulsive- and

premeditated-aggressive individuals, the current study aimed to answer the following

research questions:

1. Do impulsive-aggressive and premeditated-aggressive individuals perform

differently on measures of executive functioning known to relate to dorsolateral

prefrontal cortex functioning?

2. Do impulsive-aggressive and premeditated-aggressive individuals differ in their

recognition of emotions in faces?

3. Do impulsive-aggressive and premeditated-aggressive individuals demonstrate a

hostile attributional bias in their interpretation of neutral facial expressions or

overestimate the level of aggression in aggressive faces?

4. Do impulsive-aggressive and premeditated-aggressive individuals present with

deficits on measures of orbitofrontal cortex functioning, namely inhibition,

decision-making, and response reversal?

6

Chapter 2

Impulsive- and Premeditated-Aggression – A Review of the Literature

2.1 Aggression

Aggression can be defined as “any behaviour directed toward another

individual that is carried out with the proximate (immediate) intent to cause harm”

(Bushman & Anderson, 2001, p. 274). The aforementioned definition of aggression

encompasses a variety of behaviours, which can range from verbal to relational to

physical (Crick & Grotpeter, 1995). For example, a widely used definition of

aggression is behaviour deliberately aimed at harming people and/or objects (Dodge,

1991). In this definition, harm has implicitly been defined as hurting someone

physically. However, other forms of harm, such as psychological harm (e.g.,

humiliating), and relational harm (e.g., malevolent gossip), are just as important when

discussing the notion of aggression.

The aforementioned definition of aggression does not assume that all harmful

behaviours are aggressive, rather, there are many instances in which harmful

behaviours are prosocial. For example, the possible pain caused by a dentist to their

patient is not aggressive because the proximate intent of the dentist is to help rather

than hurt the individual. Similarly, both physical aggression in the context of self-

defence and the selective use of verbal aggression by politicians, for example, are

adaptive. Aggression, on the contrary, is problematic when it is a habitual behavioural

pattern (Bushman & Anderson, 2001).

The notion of aggression has been used to refer to a wide variety of concepts

and phenomena which has led to much confusion within the aggression literature due,

in part, to interchangeable use with other terms (Caprara et al., 1985). The related

7

theoretical constructs of anger and hostility continue to be used in lieu of aggression

and it is therefore necessary to elucidate the conceptual distinctions between these

constructs by reviewing their operational definitions. As noted above, aggression

refers to a behavioural process that includes the goal of inflicting harm to another

individual or object. In contrast, anger is conceptualised as an emotional state that can

vary in intensity, from mild annoyances to rage (Spielberger, Jacobs, Russell &

Crane, 1983). Moreover, the experience of anger lacks a specific goal (Berkowitz,

1993) and is not necessary for aggression to occur. Unlike aggression or anger,

hostility is an attitudinal or cognitive construct comprised of enduring cognitions that

involve negative interpretations of the environment. As such, once hostile attitudes

are verbally or physically expressed, they may be more appropriately labelled as

aggression. Aggressive behaviour also needs to be distinguished from antisocial

behaviour. Antisocial behaviour is defined as behaviour by which people are

disadvantaged and basic forms and norms are violated (Merk, de Castro, Koops, &

Matthys, 2005). Examples of such behaviour are lying, stealing and truancy.

Aggressive behaviour, then, is a specific form of antisocial behaviour (Kempes,

Matthys, de Vries & van Engeland, 2005).

Caprara and colleagues (Caprara et al., 1985) suggest that the use of such

concepts should be restricted to the use of two primary concepts: aggressiveness and

aggression. According to their view, aggressiveness refers to a personality

characteristic, while aggression refers to the aggressive behaviours manifested.

2.2 Distinguishing between impulsive- and premeditated-aggression

Traditionally, the study of violence and aggression has recognised the

importance of distinguishing between different types of such behaviour. Aggression

8

can be classified in a number of ways, for example, by the target of aggression (e.g.,

self-directed or directed towards another individual or object), mode of aggression

(e.g., physical or verbal, direct or indirect), or cause of aggression (e.g., medical)

(Siever, 2008). Although many individuals display more than one subtype of

aggression (Barker, Tremblay, Nagin, Vitaro, & Lacourse, 2006), and correlations

often exist among subtypes of aggression (Kempes et al., 2005), two distinct subtypes

of aggressive behaviour consistently emerge; an affective or impulsive type, and a

predatory or premeditated type. Characteristically, these two subtypes of aggression

are distinguished by several features, but primarily by the amount of behavioural

control exhibited during the incident.

Impulsive-aggression is described as a reactive or emotionally charged

aggressive response characterised by a loss of behavioural control (Barratt, 1991;

Raine et al., 1998). These aggressive acts are unplanned and spontaneous in nature

and are either unprovoked or out of proportion to the provocation. Impulsive-

aggression is usually accompanied by an agitated or irritated mood, poor modulation

of physiological arousal and loss of behavioural control (Houston, Stanford,

Villemarette-Pittman, Conklin & Helfritz, 2003). Interpersonal communication is

often non-adaptive during the agitated state and information processing appears to be

inefficient (Elliot, 1992). This subtype of aggression can result in sudden, heightened,

or inappropriate aggressive responses, and probably accounts for most societal

problems that are associated with aggression. Coccaro (1998) emphasises that while

the impulsive-aggressive individual does not necessarily have the intention to cause

harm either to themselves or to others, their behaviour nevertheless has the potential

to do so. Perpetrators of impulsive-aggression often report regret after the act,

however often lack the self control to refrain from committing such acts again

9

(Barratt, 1994). Barratt proposes that the personality traits of impulsiveness and

anger/hostility are related to most impulsive-aggressive acts.

The theoretical roots of impulsive-aggression lie in the frustration-aggression

model (Berkowitz, 1993). According to this theory, aggression is displayed as a

consequence of frustration, actual or perceived threat, and heightened arousal in the

form of anger. Aggression is displayed in reaction to aversive events with the

subjective experiences of the individual central to the situation. The subjective

experience of, for example, feeling threatened and not necessarily being threatened is

a principal concept in this theory. Frustration may not immediately lead to aggression,

but generate such emotions as anger, which can then augment the readiness to display

aggression (Merk et al., 2005).

Premeditated-aggression, on the other hand, is considered a purposeful,

controlled aggressive display that is usually instrumental in nature. These acts require

forethought and planning and are generally executed with low autonomic arousal

(Stanford, Houston, Villemarette-Pittman, & Greve, 2003b). Premeditated-aggressive

acts are carried out with a high degree of behavioural control and are directed toward

a goal, such as external reinforcers (e.g., money oriented) or intimidation (Dodge &

Coie, 1987; Hubbard, Dodge, Cillessen, Coie, & Schwartz, 2001; Vitiello, Behar,

Hunt, Stoff, & Ricciuti, 1990).

Premeditated-aggression can be understood by the principles of operant

conditioning, where the probability of aggression is increased by prior history of

subsequent reward or reinforcement (Dodge, 1991). An act of aggression occurs

because of the expectancy of the reinforcement or reward that is to follow. From this

model, one can see that the likelihood of committing an aggressive act may increase

as a function of social reinforcement emanating from an environment where gangs are

10

present or where aggression is viewed in a more positive light (Kingsbury, Lambert,

& Hendrickse, 1997). A deficit in the ability to experience or anticipate remorse or the

aversive outcomes increases the risk of premeditated-aggression, which appears to be

the case in aggressive individuals with psychopathy who have difficulty anticipating

and experiencing negative feelings of remorse or guilt (Hare, 1999).

The utility of this impulsive-premeditated classification has been indicated in

predictive validity studies (Barratt, Stanford, Felthous, & Kent, 1997a; Barratt,

Stanford, Kent & Felthous, 1997b; Heilbrun, Heilbrun, & Heilbrun, 1978; Linnoila et

al., 1983; Mungas, 1988). In a study involving male and female college students,

Barratt and colleagues (Barratt, Stanford, Dowdy, Liebman, & Kent, 1999) found

impulsive and premeditated acts of aggression are independent constructs, however,

emphasised that these two subtypes of aggressive behaviour may also coexist to

varying degrees in „normal‟ persons. It is rare in practice that the aggressive acts

exhibited by an individual can be classified as entirely premeditated or entirely

impulsive (Stanford et al., 2003b; Weinshenker & Siegel, 2002). For example, an

individual who habitually loses his temper, exhibits an irritable mood and often

responds out of proportion to the provoking stimulus might be characterised as

displaying impulsive-aggression. However, it is possible that this individual may also

demonstrate some incidences of premeditated-aggression. Conversely, an individual

whose aggressive displays are usually planned and consciously executed may

experience some instances in which he loses control of his behaviour. Thus the

proposed dimensional as apposed to categorical classification scheme permits the

behaviour to be characterised as predominantly premeditated or predominantly

impulsive, allowing for the heterogeneity that naturally occurs within an individual.

11

Although there has been some criticism of the dichotomous method of

characterising aggressive behaviour (see Bushman & Anderson, 2001; Parrott &

Giancola, 2007), a dichotomous approach is supported by a number of important

distinctions between individuals who express predominantly impulsive- or

predominantly premeditated-aggression. More specifically, researchers have found

that impulsive-aggressive individuals experience disruptions across a variety of

domains including verbal ability and intelligence, physiological reactivity, biological

function and treatment response (Barratt et al., 1997b; Coccaro, 1992). In contrast,

individuals demonstrating premeditated-aggression tend to have more circumscribed

disturbances on measures of personality (Barratt et al., 1997a; Stanford et al., 2003b).

2.2.1 Psychobiological evidence

The neurotransmitter that has received the most attention in regards to

aggressive behaviour is serotonin. Serotonin facilitates prefrontal cortical regions,

such as the orbitofrontal cortex and anterior cingulate cortex that are involved in

modulating and often suppressing the emergence of aggressive behaviours (Siever,

2008). In both humans and animals, it appears that serotonin is primarily associated

with impulsive-aggression in comparison to premeditated-aggression, and it appears

that its effects may be receptor specific (Miczek, 1987; Shaikh, De Lanerolle, &

Siegel, 1997). In humans, a relationship between decreased serotonergic function and

impulsive-aggressive behaviour has consistently been demonstrated using a number

of strategies, including central neurochemical measures (CSF 5-HIAA; Linnoila et al.,

1983; Roy, Adinoff and Linnoila, 1988; Virkkunen, De Jong, Bartko, Goodwin, &

Linnoila, 1989a), platelet binding (Kent et al., 1988), prolactin response

(Fenfluramine; Coccaro et al., 1989), pharmacological treatment (sertraline; Kavoussi,

12

Liu & Coccaro, 1994; fluoxetine; Coccaro, Kavoussi & Hauger, 1997) and regional

metabolic activity in response to serotonergic agonist (m-CPP; New et al., 2002).

Barratt, Kent, Bryant and Felthous (1991) found that phenytoin reduces the frequency

of aggressive acts. Replication studies have shown that phenytoin may reduce

incidences of impulsive-aggression, but not premeditated-aggression (Barratt et al.,

1997b; Barratt, Felthous, Kent, Liebman & Coates 2000). Such findings support the

hypothesis that impulsive- and premeditated-aggression have different underlying

biological substrates that respond differently to pharmacologic agents with specific

modes of action.

2.2.2 Psychophysiological evidence

Psychophysiological techniques also provide a practical measure of

neuropsychological functioning, however while a substantial literature exists on

autonomic correlates of antisocial behaviour (Raine, 2002a, 2000b), there are a

limited number of studies which have compared these measures in groups whose

aggression was explicitly classified according to an impulsive-premeditated scheme.

Pitts (1997) measured heart rate in aggressive children whose behaviour was

classified as reactive (impulsive) or proactive (premeditated) in nature. Those

characterised by reactive aggressive behaviour responded to a challenging task with

increasing heart rate while those characterised by proactive aggression did not.

Similarly, in a study of domestic batterers, heart rate activity was compared during

marital interaction (Jacobson & Gottman, 1998). The batterers were divided into those

whose heart rates lowered during the interaction and those whose heart rate increased.

The men exhibiting decreasing heart rates were characterised as calm, calculated,

antisocial and sadistic (premeditated). Those exhibiting increasing heart rates were

13

described as being more emotional, angry and unstable (impulsive). Finally, a more

recent study of aggressive children indicated a significant increase in skin

conductance reactivity in those rated high in reactive aggression during a laboratory-

based measure of induced anger (Hubbard et al., 2002). Again, those individuals

exhibiting primarily reactive or impulsive aggression responded differently

autonomically than those deemed more proactive or premeditated.

Electroencephalography (EEG) abnormalities are also present among those

who engage in impulsive-aggression. For example, Drake, Hietter and Pakalnis

(1992) found a greater incidence of EEG slowing in a group of patients described as

having episodic dyscontrol syndrome as compared to depressed patients and controls.

Abnormalities in P1 amplitude have been reported in impulsive-aggressive college

students (Houston & Stanford, 2001) and youths characterised by explosive

aggressive behaviours (Bars, Heyrend, Simpson & Munger, 2001). In addition, adults

classified as impulsive-aggressive exhibit decreased P1-N1-P2 latency (Houston &

Stanford, 2001), reduced P3 amplitude (Barratt et al., 1997b; Gerstle, Mathias &

Stanford 1998; Mathias & Stanford, 1999), increased P3 latency (Mathias & Stanford,

1999), and reduced amplitude and increased latency on the late positive potential, a

purported measure of emotional processing (Conklin & Stanford, 2002). These

differences reflect a number of sensory and information processing deficits specific to

impulsive-aggression, as well as preliminary evidence for emotional processing

impairment.

Volkow et al. (1995), using positron emission tomography (PET), found that

psychiatric patients with a history of repetitive, purposeless violent behaviour showed

significantly lower relative metabolic values in medial temporal and prefrontal

cortices compared to normal controls. Similarly, Raine et al. (1998) reported that

14

affective (impulsive) murderers have significantly reduced prefrontal activation when

compared to predatory (premeditated) murderers and controls.

The literature regarding premeditated-aggression though sparse is consistent.

Individuals who engage in acts of premeditated-aggression show few differences from

non-aggressive controls on psychophysiological measures, including P3 (Barratt et al.,

1997b; Stanford et al., 2003b). Stanford et al. found that the P3 latency difference did

approach significance (p = .06), suggesting a trend toward a longer P3 latency in the

premeditated-aggressive group. Such prolonged P3 latency has been linked to

increased attitudinal hostility (Bond & Surguy, 2000). Thus, the high levels of

anger/hostility evidenced in the premeditated-aggressive group may have played a

role in the latency trend observed in the sample.

2.2.3 Neuropsychological evidence

While small in number, neuropsychological studies comparing modes of

aggression have established a correlation between increased impulsive-aggression and

decreased executive functioning, while few deficits have been found in those who are

premeditated in their aggressive behaviour (Houston et al., 2003). Dolan and

Anderson (Dolan, Deakin, Roberts & Anderson, 2002) grouped male personality

disordered offenders into high and low impulsive aggressors using the Belligerence

Scale of the Special Hospital Assessment of Personality and Socialization (SHAPS:

Blackburn, 1982). They found that both impulsivity and aggression were negatively

related to executive function, and aggression was negatively related to memory

function. Similarly, Giancola, Moss, Martin, Kirisci and Tarter (1996) found that

problems in executive functioning were a predictor of reactive aggression in

adolescent boys at risk for substance abuse. Most recently, Villemarette-Pittman,

15

Stanford and Greve (2002) found that verbal deficiencies varied according to

executive demands of the task in a sample of impulsive-aggressive college students.

In the first study to compare premeditated-aggressive subjects with controls on

a variety of neuropsychological tests, Stanford et al. (2003b) found no significant

differences except for a single subscale of the Wisconsin Card Sorting Task (WCST),

where the premeditated group exhibited greater failure to maintain set than controls.

In contrast, there were pronounced differences on a range of personality measures,

including impulsivity, verbal and physical aggression, anger, hostility, psychoticism

and neuroticism. The authors concluded that the difference between the premeditated-

aggressive group and controls was a result of an impulsive personality style rather

than a significant cognitive deficit.

In summary, neuropsychological assessment has shown a clear link between

impulsive-aggressive behaviour and problems in executive functioning, while few if

any deficits have been demonstrated in premeditated-aggressive individuals.

2.2.4 Psychopathy

The concept of psychopathy has provided some utility in further distinguishing

between impulsive- and premeditated-aggression. Psychopathy refers to a

constellation of personality and behavioural characteristics marked by low baseline

arousal, dishonesty, absence of remorse, empathy, and conscience, antisocial

behaviour, and impersonal relationships (Hare, 2003). Interpersonally, they are often

described as grandiose, arrogant, callous, superficial and manipulative (Hare, 1999).

The concept of psychopathy has been operationalised by the work of Hare

(1991, 1999), and its assessment is now highly reliable and valid. There is a growing

body of research to show that psychopathic criminals engage in more premeditated

16

and predatory violence than non-psychopathic criminals (Cornell et al., 1996; Serin,

1991). Williamson, Hare and Wong (1987) found that incarcerated psychopaths had

higher rates (45.2%) than incarcerated non-psychopaths (14.6%) of committing their

crime for material gain (i.e., proactive in nature), and that non-psychopaths had higher

rates (31.7 vs. 2.4%) of emotional arousal leading to their offences (i.e., impulsive in

nature). Likewise, Cornell et al. (1996) found premeditated-aggressive offenders

could be distinguished from non-premeditated offenders by higher total psychopathy

specifically concerning: pathological lying; manipulative actions; lack of empathy;

parasitic lifestyle; irresponsibility; criminal versatility; and superficiality. The authors

concluded that “the link between psychopathy and instrumental violence supports the

distinction between instrumental and reactive violence, and raised the possibility that

the presence of instrumental violence could be an associated characteristic of

psychopathic offenders” (p. 790). Similarly, Woodworth and Porter (2002), in a study

of 125 homicide offenders, found that the great majority of homicides committed by

psychopaths were instrumental (i.e., premeditated), whereas only 48.4% of homicides

committed by non-psychopaths were instrumental. Meloy (1988) theorised that a

predisposition to engage in premeditated violence in psychopaths would be due to

their low levels of autonomic arousal and reactivity, their disidentification with the

victim, their emotional detachment and their lack of empathy.

Similar differences have been found in children. A growing body of evidence

indicates that children with conduct problems represent a group that can be further

divided into discrete subtypes based on the presence of callous and unemotional traits

(Frick, O‟Brien, Wootton & McBurnett, 1994; Frick & Ellis, 1999). Callous and

unemotional traits are similar to psychopathic traits such as lack of guilt and empathy,

superficial charm, and constricted emotion. In a sample of children with conduct

17

problems, Christian, Frick, Hill and Tyler (1997) found that a subgroup of children

exhibiting symptoms of Oppositional Defiant Disorder (ODD) and Conduct Disorder

(CD) and callous and unemotional traits differed from those without such traits in the

number and variety of conduct problems. Pardini, Lochman and Frick (2003) found

that the presence of callous and unemotional traits in adjudicated juveniles was

associated with the use of aggression to obtain rewards and dominate (i.e.,

premeditated). Frick et al. (2003), in a sample of non-referred children, found that

children demonstrating conduct problems and callous and unemotional traits were

more likely to demonstrate high levels of proactive aggression than those without

these traits, whose aggression was predominantly impulsive. It thus appears that it is

the presence of callous and unemotional traits that distinguish this subgroup and its

associated problems.

2.3 Importance of distinguishing between impulsive- and premeditated

aggression

There is evidence in both children and adults that impulsive- and

premeditated- aggression are distinguishable forms of aggressive behaviour with

important clinical implications. For example, longitudinal studies in children and

adolescents rated on reactive and proactive aggressive behaviour have shown that

proactive, but not reactive, aggression predicts later delinquent behaviour (Vitaro,

Gendreau, Tremblay & Oligny, 1998). Nouvion and colleagues found that proactive

aggressive subjects had a greater incidence of personality disorders, including CD and

APD, compared to reactive aggressive and non-aggressive control subjects (Nouvion,

Cherek, Lane, Tcheremissine & Lieving, 2007).

18

These findings are in line with previous studies that have found proactive, but

not reactive, aggression to be predictive of ODD, CD and externalising problems

(e.g., Conner, Steingard, Anderson & Melloni, 2003; Pulkkinen, 1996; Vitaro et al.,

1998). Given that individuals diagnosed with APD must have evidence of CD before

the age of 15, early onset behavioural problems leading to adult antisocial behaviour

was prevalent in the premeditated group. This provides evidence that premeditated-

aggressive individuals may have increased personality psychopathology and be at

increased risk for early aggressive and antisocial behaviours relative to impulsive-

aggressive or non-aggressive individuals. The fact that CD was a distinguishable

factor between groups is of significance due to the fact that CD is stable over time

(Bassarath, 2001a; Kazdin, 2000), and is associated with criminal behaviour and

substance abuse (Hser, Grella, Collins & Teruya, 2003; Mueser et al., 2006;

Tcheremissine & Lieving, 2006).

Heilbrun et al. (1978) found that murderers whose violence was classified as

impulsive were more likely to fail on parole that those whose murders were

premeditated in nature. In a pharmacological treatment study comparing aggressive

subtypes, inmates whose aggressive behaviour was classified as impulsive showed

significant reductions in the frequency and intensity of aggressive acts, normalisation

of event-related potentials (P3) and improvement in mood state measures during a six

week trial of anticonvulsant phenytoin (Dilantin) when compared to placebo (Barratt

et al., 1997a). Inmates whose aggressive behaviour was classified as premeditated in

nature showed no improvement during the same trial. Similarly, Malone et al. (1998)

examined the effect of lithium carbonate in CD children whose repeated aggressive

behaviour was categorised as either predatory or affective in nature. Treatment

response was associated with affective rather than predatory aggressive behaviour.

19

While distinguishing between these forms of aggression can not only lead to a

better theoretical understanding of aggression (Coie & Dodge, 1998; Poulin & Boivin,

2000; Vitiello & Stoff, 1997), it can also to better prognostication. Such a distinction

is also assumed to lead to the development of more specific interventions and

treatment, which may then prove more effective than interventions aimed at

aggression in general (McAdams, 2002; Vitello & Stoff, 1997).

20

Chapter 3

Prefrontal Cortex & Aggression

Research into the antecedents of violence and aggression indicates that there

are many factors which contribute to the development of these behaviours. It is

important to note that while there are general predictors of violent and aggressive

behaviour, no single theory can account for causation in all situations. It is accepted

that the causes of aggression are multi-faceted and that neurological deficit may be a

factor in only a small percentage of those who demonstrate such behaviour. However,

given that aggression – like any behaviour – ultimately derives from the normal and

abnormal operations of the brain, closer examination of the aspects of brain structure

and function relevant to aggressive behaviour are required.

Numerous studies, in both animals and humans, have supported an association

between abnormalities in brain function and aggressive and violent behaviour (Filley

et al., 2001; Golden, Jackson, Peterson-Rohne & Gontkovsky, 1996; Krakowski,

2003). Case studies of patients with neurological disorders or those who have suffered

traumatic brain injury provide provocative insights into which brain regions, when

damaged, might predispose to irresponsible, aggressive behaviour.

Psychophysiological and neuropsychological assessments have also demonstrated that

violent and/or aggressive individuals have lower brain functioning than controls,

including lower verbal ability and diminished executive functioning (Barratt et al.,

1997b; Dolan & Park, 2002; Hoaken et al., 2007).

The availability of new functional and structural neuroimaging techniques,

such as PET, magnetic resonance imaging (MRI) and functional MRI (fMRI), has

broadened our understanding of the neural circuitry that underlies aggressive and

21

antisocial behaviours. More specifically, as reviewed by Davidson and colleagues, a

circuit that includes several regions of the prefrontal cortex, the amygdala,

hippocampus, hypothalamus, anterior cingulate cortex, ventral striatum, and other

interconnected structures has been implicated in various aspects of emotion regulation

and affective style (Davidson & Irwin, 1999; Davidson, Jackson & Kalin, 2000a).

Emotion regulation includes those processes which amplify, attenuate, or maintain an

emotion, and thus incorporates the expression of aggressive behaviours. Related to

this is evidence which suggests that individuals who are vulnerable to faulty

regulation of negative emotion may be at increased risk for aggression and/or violent

behaviour (Davidson, Putnam & Larson, 2000b).

Prefrontal cortical dysfunction has been implicated as a possible anatomical

correlate of aggressive behaviour (Convitt et al., 1996; Critchley et al., 2000; Pietrini,

Guazzelli, Basso, Jaffe & Grafman, 2000; Raine et al., 1994). Both cognitive and

behavioural similarities have been noted between individuals who have frontal lobe

damage and those who show characteristics of antisocial behaviour (Price, Daffner,

Stowe & Mesulam, 1990). For example, Eslinger and Damasio (1985) noted that

damage to the prefrontal area is associated with heightened aggression, emotional

outbursts, disorganisation, and risk-taking behaviour. This observation has been a

major impetus for the research on neuropsychological abnormalities in antisocial

individuals. In particular, damage to the prefrontal regions of the brain and the

resulting impairment in executive functions is considered to be associated with

increased aggressive and antisocial behaviour (Brower & Price, 2001; Tateno, Jorge

& Robinson, 2003).

22

3.1.1 The prefrontal cortex

The frontal cortex encompasses the brain areas anterior to the central sulcus

and comprises approximately one third of the cerebral cortex. The frontal cortex can

be divided into three principle regions: the primary motor cortex, the prefrontal

cortex, and the limbic cortex (Duke & Kaszniak, 2000). The prefrontal cortex refers to

the most anterior regions of the frontal lobes and it is functionally and anatomically

heterogeneous (Fuster, 2001). The prefrontal cortex has a rich supply of connections

with other neural regions. Cortically, it is connected with association cortex in the

temporal, parietal and occipital lobes, and subcortically with the hippocampus,

amygdala, thalamus, hypothalamus, subthalamus, septum, striatum, pons, and

mesencephalon (Fuster, 2001; Pandya & Barnes, 1987). Given that the prefrontal

cortex is connected to more brain regions that any other cortical region, its position

allows the integration of information processed at lower levels, including input from

the limbic circuits, as well as being the major target of the basal ganglia-

thalamocortical circuits (Royall et al., 2002).

The prefrontal cortex, along with its underlying subcortical regions, is

extensively interconnected with the major sensory and motor systems of the brain.

Connections from the posterior cortical areas, particularly areas of multimodal

convergence, bring information regarding the external environment. Subcortical

pathways, including the amygdala, hippocampus, midbrain area, and thalamus, bring

details about internal states (Duke & Kaszniak, 2000).

The prefrontal cortex has direct neural projections from a variety of limbic

structures that are directly linked to the amygdala. There are input projections from

the thalamus to the prefrontal cortex, and these connections contain information

arising from the temporal cortex and the amygdala. Direct reciprocal connections

23

from the prefrontal cortex to the amygdala have also been identified (Afifi &

Bergman, 1998). Output projections to the amygdala are both excitory and inhibitory

in nature. However, damage to the prefrontal cortex results in an overactivation of the

amygdala, suggesting that the effect of the prefrontal cortex on the amygdala is

predominantly inhibitory (Gerwitz, Falls & Davis, 1997; Morgan, Romanski &

LeDoux, 1993). With regard to the functionality of this connection, lesions to the

prefrontal cortex in rats reduce the prefrontal inhibitory action on the amygdala,

resulting in an increased difficulty in the extinction of aversive responses (Morgan et

al., 1993), as well as impairing the ability to anticipate future negative consequences

(Bechara, Tranel, Damasio & Damasio, 1996). Therefore, it appears that the

prefrontal cortex plays an important role in regulating the acquisition of new

responses, and the extinction of aversive responses (Lopez, Vazquez & Olson, 2004).

The prefrontal cortex also plays a central role in many aspects of social

cognition (Rilling et al., 2002), including perspective taking (Frith & Frith, 1999), and

also in the regulation of emotions such as aggression (see Blair, 2004 for review).

Early descriptions of frontal lobe syndromes arose from several 19th century

investigators, described in a number of reviews (Damasio et al., 1994; Macmillan,

2002; Tranel, Anderson & Benton, 1994), highlighting the changes displayed in social

behaviour, personality, and emotional regulation that occurred after frontal lobe

pathology. Subsequent investigators continued to elaborate on the nature and extent of

these deficits, their causes and management (Miller & Cummings, 1999; Stuss &

Knight, 2002), firmly establishing a vital role for the frontal lobes, particularly the

prefrontal cortex, in such processes.

In an overview of neuroanatomy and neuropathology, Stuss and Benson

(1984) described six specific manifestations of prefrontal damage: (1) inability to use

24

knowledge to regulate behaviour; (2) impaired ability to handle sequential behaviour;

(3) impaired ability to establish or change a mental set; (4) impaired ability to

maintain a mental set; (5) impaired ability to monitor personal behaviour; and (6)

attitudes of apathy.

3.2 Prefrontal divisions

For clinical purposes, the prefrontal cortex can be divided into three distinct

neuroanatomical regions: 1) dorsolateral prefrontal cortex (Brodmann‟s areas 9, 10,

46); 2) medial prefrontal cortex (including the functionally related anterior cingulate

cortex and Brodmann‟s area 24); and 3) orbital prefrontal cortex (Brodmann‟s areas

11 and 12), corresponding to the most inferior and ventral parts of the prefrontal

cortex (behind the eyes, or orbits). Both medial prefrontal and orbitofrontal are part of

a frontostriatal circuit that has strong connections to the amygdala and other parts of

the limbic system. Consequently, these regions are anatomically well suited for the

integration of affective and non-affective information, and for the regulation of

appetitive/motivated responses. Functionally, these regions are often considered

together, as when researchers focus on effects of damage to ventromedial prefrontal

cortex (Happaney et al., 2004).

The prefrontal cortex is a heterogeneous region of the brain and the three

principal frontal-subcortical circuits are involved in cognitive, emotional, and

motivational processes. The primary focus of the current research will be on the roles

of the dorsolateral and orbital divisions of the prefrontal cortex, which manifest quite

distinct anatomical and functional properties (Fuster, 1989; Stuss & Benson, 1986).

The dorsolateral prefrontal cortex projects primarily to the dorsolateral head

of the caudate nucleus, which receives input from the posterior parietal cortex and

25

premotor areas. The dorsolateral circuit then connects to the dorsolateral part of the

globus pallidus and rostral substantia nigra reticulate, and continues to the

parvocellular area of the medial dorsal and ventral anterior portions of the thalamus.

Projections from the thalamus back to the dorsolateral prefrontal circuit close the

circuit (Cummings, 1993).

Functionally, the high-level cognitive abilities mediated by the dorsolateral

prefrontal cortex and its connections are those referred to as „executive functions‟,

including cognitive flexibility, temporal ordering of events, planning, monitoring and

inhibiting pre-programmed behaviour, set-shifting, working memory and concept

formation (Smith & Jonides, 1999). According to Cummings (1995), dysfunction in

the dorsolateral prefrontal circuit is associated with circuit-specific problems

including decreased verbal fluency, perseveration, difficulty shifting set, poor

recall/retrieval of information, reduced mental control, limited abstraction ability, and

poor response inhibition. However, while patients with lesions restricted to this region

are concrete and perseverative and show impairments in reasoning and mental

flexibility (Benton, 1986), they typically demonstrate intact perception, calculation,

language abilities and storage of memories (Duke & Kaszniak, 2000).

The orbitofrontal cortex occupies the ventral region of the prefrontal cortex

(Kringelbach & Rolls, 2004), which is reciprocally connected with the amygdala

(Ghashghaei & Barbas, 2002). The orbitofrontal cortex projects to the ventromedial

caudate nucleus, which receives input from other cortical association areas and

brainstem regions, and has open interconnections with the dorsolateral prefrontal

cortex, the temporal pole, and the amygdala (Davis & Whalen, 2001). The

orbitofrontal cortex contains the secondary taste cortex, in which the reward value of

taste is represented. It also contains the secondary and tertiary olfactory cortical areas,

26

in which the identity and also the reward value of odours are represented. The

orbitofrontal cortex also receives information about the sight of objects from the

temporal lobe cortical visual areas (Rolls, 1999).

The orbitofrontal-subcortical circuit is said to underlie social behaviour and

appears to play a critical role in the representation of the reward value of a stimulus

and the way in which this representation guides goal-directed behaviour (Rolls, 1999).

Lesions specific to the circuit have been found to result in marked changes in

personality, including disinhibition, impulsivity, and antisocial behaviour, and

irritability and lability are often prominent (Cummings, 1995). Some of the changes

may be related to difficulty in the learning and reversal of stimulus-reinforcement

associations, and thus the correction of behavioural responses when they are no longer

appropriate due to changes in reinforcement contingencies (Rolls, 2004; Hornak et al.,

2004). Indeed, investigations in macaques have shown that lesions to the orbitofrontal

cortex impair reversal learning (Dias et al., 1996a). Consistent with this, the

orbitofrontal cortex is activated by monetary rewards and punishments, and the

magnitude of the reinforcers (O‟Doherty, Kringelbach, Rolls, Hornak & Andrews,

2001). The visual input to neurons in the orbitofrontal cortex is in many cases the

reinforcement association of visual stimuli, one of which is information about faces.

Such facial stimuli convey information that is important in social reinforcement

(Rolls, 2004).

The medial circuit, begins in the anterior cingulate and projects to the nucleus

accumbens. The anterior cingulate has interconnections with dorsolateral prefrontal

cortex and the amygdala, and it also receives input from the ventral tegmental area

(Duke & Kaszniak, 2000). The medial frontal-subcortical circuit is involved in

27

motivation. Lesions to this region often produce apathy, lack of motivation, decreased

social interaction, and psychomotor retardation (Sbordone, 2000).

The ventromedial prefrontal region includes the medial and varying sectors of

the lateral orbitofrontal cortex, encompassing Brodmann‟s areas 25, lower 24, 32, and

medial aspect of 11, 12, and 10, and the white matter subjacent to all of these areas

(Bechara, 2004). Patients with bilateral lesions of the ventromedial cortex develop

severe impairments in personal and social decision-making, in spite of otherwise

largely preserved intellectual abilities. Following damage to this region of the

prefrontal cortex, patients develop difficulties in daily and future planning, and

difficulties in choosing friends and activities (Bechara, Damasio & Damasio, 2000a;

Bechara, Tranel & Damasio, 2002).

The identification of these adjacent circuits provides insight as to the

similarities of behavioural changes caused by lesions to different brain regions. Whilst

focal lesions to the areas of the prefrontal cortex have led to what have been labelled

“frontal lobe syndrome”, the involvement of multiple circuits in subcortical lesions

has resulted in variable behavioural manifestations (Cummings, 1995). For example,

studies of lesions to the globus pallidus have described patients with marked changes

in personality and reduced activity levels with memory and executive function

deficits, but with normal intelligence and language abilities (e.g., Strub, 1989).

In summary, the frontal–subcortical circuits are extensively connected to each

other at the level of the frontal lobes. The circuits are discrete in subcortical regions.

The dorsolateral circuit, because of its neuroanatomy, is uniquely able to integrate

information from all three frontal–subcortical circuits. Here, the integrated

information from the external world and the cognitive and emotional states of the

individual can be used in the production of social behaviour.

28

3.3 Prefrontal dysfunction and aggression

Frontal lobe dysfunction in particular, has been invoked to explain the actions

of individuals convicted of violent crimes, who appear to fail to inhibit impulsive,

trivially motivated, or habitual aggression. Case studies as far back as 1935 have

reported the onset of antisocial personality traits after frontal lobe injury (Blumer &

Benson, 1975). Such cases typically involve damage to the orbitofrontal cortex, which

clinical observations have associated with poor impulse control, explosive aggressive

outbursts, inappropriate verbal lewdness, jocularity, and lack of interpersonal

sensitivity (Duffy & Campbell, 1994). This dysregulation of affect and behaviour may

occur while cognitive, motor, and sensory functioning remains relatively intact

(Mesulam, 1986).

3.3.1 Lesion studies

Research on individuals who have suffered traumatic brain injury is of key

importance in investigating the neural substrates of aggressive behaviour. The critical

role of the prefrontal cortex in aggressive behaviour was initially recognised by case

reports that prefrontal brain lesions could result in the emergence of antisocial

behaviours or psychopathic traits in previously normal subjects (Damasio, Tranel &

Damasio, 1990). A prime example of this disinhibition is found in the often cited case

of Phineas Gage, a dependable and responsible stable railroad worker who was

injured by a tamping rod that penetrated his skull through his orbital frontal cortex.

After the accident he became irresponsible and impulsive, despite preserved general

cognitive and motor skills (Damasio et al., 1994).

McAllister and Price (1987) evaluated 20 psychiatric patients with diverse

types of frontal cortical pathology as identified by computed tomography scans and

29

EEG. Results indicated that 60% of the patients displayed disinhibited behaviour with

affective lability and 10% displayed violent outbursts. However, the results of this

study are difficult to interpret given that, in addition to having a frontal lobe

pathology, all of the patients had at least one psychiatric diagnosis, and the exact

neuroanatomical location of the pathology for each patient was not reported. In

another study, Heinrichs (1989) found that a frontal cortical lesion was the best

predictor of violent behaviour in a sample of 45 neuropsychiatric patients. Again,

many of the patients in this study had other psychiatric diagnoses and the exact

anatomical locations of the frontal neuropathologies were not specified.

Further data from neurological case reports have provided much useful

information regarding the relationship between prefrontal cortical functioning and

aggression. Thompson (1970) reported a case of a 33-year-old male with a history of

violent behaviour subsequent to a head injury at the age of 12. A

pneumencephalography revealed bilateral cortical atrophy in the prefrontal regions.

Price et al. (1990) studied the adult behaviour patterns of two patients who acquired

brain damage during childhood. While both patients developed relatively normally

until the damage was sustained, following the damage these patients displayed an

inability to respond to punishment or delay gratification, irresponsibility, sexual

promiscuity, grand larceny, drug involvement, angry outbursts, arson, suspected rape,

and physical violence. Although the exact location of the lesions in both cases is

equivocal, neurological and neuropsychological examination indicated bilateral

lesions in the prefrontal cortex.

Boone et al. (1988) described a 13-year-old female suffering from partial

complex seizures localised primarily in the prefrontal cortex. Six weeks before being

admitted to hospital she began exhibiting prominent behavioural changes consisting

30

of bizarre speech, sexual disinhibition, disobeying parental orders, and verbal and

physical aggression. An EEG revealed activity in the frontal lobes and

neuropsychological tests demonstrated deficits on prefrontal tests involving attention,

alternation between tasks, performance on mazes, response inhibition, and

distractibility. In a similar case, Eslinger and Damasio (1985) noted personality

changes in patient EVR subsequent to surgical ablation of the orbital and mesial areas

of the prefrontal cortex. Following the surgery, while his level of intelligence was

above average, EVR began to engage in what the authors termed „sociopathic

behaviour‟, including difficulties in decision-making, adjustment problems, poor

judgement, and employment problems. The patient performed well on prefrontal

cortical tests such as the WCST which the authors attributed to the fact that the

dorsolateral prefrontal regions and superior mesial regions were left undamaged.

Meyers, Berman, Scheibe and Hayman (1992) noted similar behavioural

sequelae involving disinhibition, poor judgement, and irresponsibility subsequent to

surgical damage to the left orbital prefrontal cortex in a 33-year-old male.

Interestingly, this patient performed in the above average range on prefrontal cortical

tests such as the WCST; however this is again likely due to the preservation of the

dorsolateral prefrontal cortex. These findings reflect those of Phineas Gage described

earlier, who subsequent to his injury, was described as being untrustworthy,

irresponsibly, and disrespectful. Again, the majority of the neural damage was located

in the orbital and mesial prefrontal regions, whereas the dorsolateral area was found to

be spared (Damasio et al., 1994).

Of the eight studies reviewed in the preceding section, six document a

relationship between prefrontal cortical pathology and aggressive behaviour.

However, none of these reports specified the exact location of the neuropathology

31

within the frontal lobes. As a result, these data do not provide evidence to implicate

more specifically the dorsolateral or orbital prefrontal regions in the expression of

aggressive behaviour. The remaining two reports indicate that their patients had

lesions in the orbital prefrontal cortex and not the dorsolateral area.

Other case studies of patients who have sustained damage to the orbitofrontal

region, such as EVR, resemble Gage in manifesting a behavioural profile that has

been referred to as „acquired sociopathy‟ (Saver & Damasio, 1991; Tranel, 1994;

Meyers et al., 1992; Blair & Cipolotti, 2000). Damasio et al. (1994) describe acquired

sociopathy as a reactive, emotionally driven violence toward a person that is related to

emotional inhibitory dyscontrol. Although showing minimal impairments on standard

neuropsychological tests of intelligence and executive functions, these patients

display marked deficits in real life tasks involving judgement, awareness of socially

appropriate conduct, and the capacity to assess future consequences (Bechara et al.,

2000b).

Blair and Cipolotti (2000) reported on JS who sustained damage to the

orbitofrontal cortex and some damage to the left amygdala. Premorbidly, JS was

described as being a quiet, withdrawn person who was never aggressive. Following

the damage, JS showed unpredictable, impulsive-aggression and violence, and

demonstrated deficits in the recognition of facial expression, particularly in the

recognition of anger and disgust. He also produced significantly lower skin

conductance responses (SCR) to the anger and disgust expressions compared with

comparison groups.

Further clinical data demonstrate that lesions in the orbitofrontal cortex and

adjacent prefrontal regions produce syndromes characterised by impulsivity and

aggression. Anderson et al. (1999) reported on two individuals, tested in their

32

twenties, who suffered early damage to orbital and lateral sectors of the prefrontal

cortex. Both exhibited a significant deficit in moral reasoning, a history of verbal and

physical aggression, and intermittent, explosive bursts of anger. A further study of

two adults who sustained frontal lobe injury in childhood suggests that early damage

to orbitofrontal regions may lead to a “comportmental learning disability” that closely

resembles sociopathy and includes a diminished capacity to inhibit violence (Price et

al., 1990).

Further evidence implicating the orbitofrontal regions comes from a large

retrospective study of Vietnam veterans with penetrating head injuries, which found

that ventromedial frontal and orbitofrontal lesions, as assessed by computed

tomography scans, specifically increased the risk of aggressive and violent behaviour.

(Grafman et al.,1996). Data have also been reported showing higher rates of antisocial

behaviour (including stealing, physical assault and sexual comments or advances) in

patients with frontotemporal dementia, even when compared with equally cognitively

impaired patients with Alzheimer‟s disease (Miller, Darby, Benson, Cummings, &

Miller, 1997; Stip, 1995).

3.3.2 Neuroimaging and clinical neurological findings

Brain imaging studies are now beginning to confirm the role of the prefrontal

cortex in modulating and controlling violence in humans. Reviews of brain imaging

studies of violent and psychopathic populations completed by Raine (1993), Mills and

Raine (1994), Raine and Buchsbaum (1996), Henry and Moffitt (1997), and Bufkin

and Luttrell (2005), while showing some variability across studies, concur in

indicating that violent offenders have functional deficits in the anterior regions of the

brain, particularly the frontal region.

33

Most recently, in their review of 17 neuroimaging studies, Bufkin and Luttrell

(2005) found that the areas associated with aggressive and/or violent behaviour,

particularly impulsive acts, are located in the prefrontal cortex and the medial

temporal regions. Of the 17 studies reviewed, 14 specifically examined possible links

between frontal lobe pathology and aggressive and/or violent behaviour. In the 10

single photon emission computed tomography (SPECT) and PET studies, 100%

reported deficits in either prefrontal (8 of 10 studies) or frontal (2 of 10 studies)

functioning in aggressive, violent and/or antisocial groups compared to non-

aggressive patients or healthy controls. Analyses of specific regions in the prefrontal

cortex revealed that individuals who were aggressive and/or violent had significantly

lower prefrontal activity in the orbitofrontal cortex (4 of 10 studies), anterior medial

cortex (2 of 10 studies) and/or superior frontal cortex (1 of 10 studies). In the four

MRI studies, half reported decreased grey matter volume in prefrontal or frontal

regions, and 25% reported non-specific white matter abnormalities, not localised to

the frontal cortex.

Initial studies demonstrate anterior brain dysfunction in individuals with a

history of violence. Goyer et al. (1994), using PET in an auditory activation condition,

showed that an increased number of aggressive impulsive acts were associated with

reduced glucose in the anterior medial and left anterior orbitofrontal frontal cortex of

17 personality disordered patients. Volkow and colleagues in two PET studies which

compared forensic psychiatric patients with normal controls, documented decreased

frontal cortical blood flow or metabolism associated with „repetitive‟ and

„purposeless‟ violent behaviour (Volkow & Tancredi, 1987; Volkow et al., 1995).

Søderstrom et al. (2000) using SPECT, found reduced blood flow in both frontal and

temporal lobes of 21 individuals convicted of impulsive violent crimes.

34

In a PET study evaluating responses to the probe metachlorophenylpiperazine,

decrements in the lateral, medial and orbital frontal cortices were found at baseline in

men with a history of physical aggression and in the orbital frontal cortex for both

men and women with a history of physical aggression (New et al., 2009).

Furthermore, a series of studies demonstrated that while normal subjects show

increased relative glucose metabolism in orbitofrontal cortex and anterior cingulate

gyrus following acute serotonergic stimulation, impulsive-aggressive personality

disordered patients show decreased relative metabolism in this area (New et al., 2002;

Siever et al., 1999; Soloff et al., 2003). These studies suggest that orbitofrontal and

adjacent regions may exert an inhibitory influence on aggression, perhaps through a

serotonergic mechanism.

Interictal episodes of impulsive aggression have also been observed in patients

with temporal lobe epilepsy. Such patients who display episodes of impulsive

aggression have a highly significant reduction (approximately 17%) in left prefrontal

gray matter compared with temporal lobe epilepsy patients with no history of

aggression or controls (Woermann et al., 2000).

One particular set of studies was undertaken by Raine and colleagues on a

heterogeneous group of suspected murderers pleading not guilty by reason of insanity.

In a first study Raine et al. (1994) conducted a PET study on 22 subjects accused of

murder and 22 matched controls. The offender group had significantly lower glucose

metabolism in both medial and lateral prefrontal cortex relative to the controls. Raine,

Buchsbaum and LaCasse, (1997) in a study of 41 murderers found hypoactivation in

prefrontal territories including lateral and medial regions of the prefrontal cortex, as

well as activation in the right amygdala, compared with age- and sex-matched

controls. In a subsequent reanalysis of these data, murderers were classified as those

35

who commit planned, predatory murder or those who committed affective, impulsive

murder. The impulsive murderers showed reductions in lateral prefrontal metabolism

compared with controls, whereas the predatory group did not (Raine et al., 1998).

Findings from a more recent structural MRI study indicated that individuals with a

diagnosis of APD recruited from the community showed an 11% reduction in the

volume of gray matter in the prefrontal cortex, compared to both normal controls and

a substance dependence control group (Raine et al., 2000).

A number of studies have also found abnormal frontal EEG activity, as well as

diminished frontal event related potentials (ERP), correlating with antisocial

personality disorder or histories of aggression (Bauer, O‟Connor & Hesselbrock,

1994; Bernat, Hall, Steffen, & Patrick, 2007; Finn, Ramsey & Earleywine, 2000;

O‟Connor, Bauer, Tasman & Hesselbrock, 1994).

3.4 The specific roles of the orbitofrontal and dorsolateral prefrontal cortex in

aggression

The cumulative evidence from the studies reviewed above indicates that

clinically significant frontal lobe dysfunction is associated with aggressive behaviour.

Subjects with both traumatic brain injury and neurodegenerative disorders primarily

involving the prefrontal cortex display increased rates of aggressive and antisocial

behaviour compared to subjects who have no, or non-frontal brain injury. Studies

employing neuropsychological testing, neurological examination, EEG, and

neuroimaging have also found evidence for increased rates of prefrontal network

dysfunction among aggressive and antisocial subjects (see Brower & Price, 2001;

Bufkin & Luttrell, 2005 for reviews).

36

Unfortunately, however, the above studies have placed little emphasis on

considering the separable regions of frontal cortex. Data based on acquired damage to

the prefrontal cortex